Privacy Policy
Effective: June 11, 2026Last updated: August 4, 2026
Stack Stitch (the "Service") is operated by StackStitch ("we", "us") — currently its founding team, based in Colombia; a legal entity is being formed and this policy will be updated to name it once incorporated. This policy is governed by the laws of Colombia (Ley 1581 de 2012 — Habeas Data). Responsible party: StackStitch — hello@stackstitch.dev. The legal entity and responsible-party details will be updated once incorporated and reviewed by counsel.
1. Summary (plain language)
Stack Stitch is a macOS desktop app and hosted service that connects to Slack, GitHub, Gmail, Notion, Granola, Jira, Sentry, and Stack Stitch Calls, then surfaces relevant context with its sources attached. To provide the Service, we process content you authorize and use providers for AI generation, transcription, hosting, billing, analytics, and diagnostics. We do not sell personal data. You choose what to connect and when to record a call. Call recording and participant consent remain your responsibility (see §7).
2. Information we collect
a. Account & identity. When you sign in (via Google / Firebase Authentication) we receive your email, name, and an authentication token. We store a user record and authentication state.
b. Data from connected sources. Depending on the connector and authorization you choose, we may access Slack messages and threads; GitHub repositories, pull requests, reviews, checks, and issues; Gmail messages and threads; Notion pages and comments; Granola meeting notes and transcripts; Jira issues and comments; and Sentry project issues, alerts, and regressions. This content may include information about third parties. See §8.
c. Calls, audio & participant labels. Stack Stitch can detect a supported calling app and show a prompt, or you can start recording manually. Capture begins only after you choose Record. We may process microphone and system audio; transcripts, timestamps and speaker labels; call summaries and action items; and limited frames of the focused meeting-client window, sampled periodically and processed on-device to read visible participant-name labels. We retain extracted name samples, not a continuous screen video. We may also store speaker embeddings or "voiceprints" for consistent speaker labels across calls; these are not used for authentication.
d. Content & derived data. From connected content we create and store source artifacts, proactive notifications, correlated context, discussion threads, call insights, and vector embeddings used for retrieval.
e. Billing data. Payments are processed by our Merchant of Record, Polar (which uses Stripe). We do not receive or store your full card details. We store your subscription status and provider identifiers (e.g. customer/subscription IDs, invoice metadata).
f. Product analytics. If enabled, a limited event catalog is sent to PostHog and may temporarily be sent to Google Analytics for migration validation. Signed-in PostHog events use a salted, one-way user hash rather than your email, name, or raw account ID; person-profile processing is disabled and we do not create PostHog People profiles. You can disable product analytics in the app.
g. Reliability diagnostics. Production builds may send scrubbed crash and error diagnostics to Sentry independently of product analytics. Sentry is configured without default PII collection; account correlation uses a salted hash, and common sensitive fields and patterns are removed or masked.
3. How we use your information
- To provide the Service: ingest connected content, generate source-grounded notifications, process calls, and power retrieval, discussions, and call insights.
- To authenticate you and keep your account secure.
- To process billing (via Polar) and manage trials/subscriptions.
- To maintain, debug and improve reliability and quality.
- To communicate with you about the Service.
We process data to provide the Service you request, based on your authorization or consent where applicable, and for our legitimate interests in operating and securing the product. We do not sell personal data. We do not use connected-source content, call audio, or transcripts to train our own models. Where provider controls permit, we configure providers not to train on customer content. We may use aggregated or de-identified data to operate and improve the Service.
4. AI processing
We send selected content excerpts to OpenRouter and model providers routed through it to produce notifications, summaries, discussions, and call insights. Some embeddings run locally; configured hosted embedding calls may use OpenRouter. Call audio is sent to Speechmatics for transcription and speaker diarization. Speechmatics may return transcripts, timestamps, speaker labels, and voiceprint data. According to its current batch-service documentation, submitted job data is retained for up to seven days. AI and transcription output may be inaccurate, incomplete, or misattribute a speaker — see the Terms.
5. Sub-processors
We rely on the following third parties to operate the Service. Each processes data only to provide its function:
| Sub-processor | Purpose | Data involved |
|---|---|---|
| Google Firebase | Authentication / identity | Email, name, auth tokens |
| OpenRouter and routed model providers | AI generation and, when configured, embeddings | Selected connected-content and transcript excerpts |
| Speechmatics | Call transcription, diarization, and speaker attribution | Call audio, transcript data, speaker labels/voiceprints |
| Polar (Merchant of Record) + Stripe | Payments, invoices, tax | Billing/payment data (handled by them) |
| MongoDB Atlas | Primary hosted database | Account, connected-source, call, and derived product data |
| DigitalOcean | Application and call-service hosting, audio storage, self-hosted vector index | Data processed by the Service in transit and at rest |
| PostHog | Product analytics | Pseudonymous usage events; person profiles disabled |
| Google Analytics (GA4) | Temporary analytics comparison/fallback | Limited usage events |
| Sentry | Error, crash, reliability, and security diagnostics | Scrubbed diagnostic events and salted account hash |
The Service also communicates with each third-party source you explicitly connect; those sources operate under their own terms and privacy policies. We will provide notice of material changes to this list.
6. Where your data is processed / international transfers
The Service uses providers and infrastructure located outside Colombia, including in the United States and other jurisdictions. This may constitute an international transfer or transmission of personal data under Ley 1581. By authorizing the relevant features and using the Service, you authorize the processing needed to provide them, subject to applicable law.
7. Call recording & consent (read this carefully)
Stack Stitch may detect that a supported calling app is active and show a prompt, but detection alone does not begin capture. You can also start a recording manually. Capture starts only after you choose Record. You are responsible for complying with all applicable recording, wiretapping, employment, and privacy laws, which may require the consent of every participant. By recording, you represent that you have the required authority and consent. Call audio, transcripts, visible participant labels, and speaker voiceprints may be personal or sensitive data.
8. Third-party data & your authority
When you connect a source or capture a call, the data may include information about other people. By connecting sources and enabling features, you represent that you have the authority and right to do so (including any authorization from your employer/organization) and to allow us to process that data on your behalf.
Our role differs by data category: for customer-controlled workspace and call content, you or your organization generally act as controller and Stack Stitch processes the data on your instructions. For account, access, billing, analytics, diagnostics, security, support, and our own operational decisions, Stack Stitch acts as controller under this policy.
9. Data retention
| Data | Retention |
|---|---|
| Account + connected/derived content | While the account is active or as needed to provide the Service; verified deletion requests are subject to legal and backup requirements |
| After source disconnect or account deletion | Target removal from active systems within 30 days; backups roll off within 90 days |
| Desktop call intermediates | Removed after successful handoff/processing; capture failures may be retained until retry, dismissal/deletion, or operational cleanup |
| Server-side call audio | Deleted after a transcript passes the quality gate; failed or suspect sessions retain audio so the user can retry, until deletion or operational cleanup |
| Speechmatics batch job data | Up to 7 days under the provider's current documented retention |
| Anonymous speaker voiceprints | Expire 30 days after last seen |
| Confirmed/owner speaker voiceprints | While needed for speaker labeling, generally while the account remains active or until deletion is requested |
| Operational logs | 30 days |
| Security / audit logs | 90 days |
| Minimal billing records | While account active + as required by law (Polar retains payment/tax records) |
Early-access retention note: failed or suspect call audio is currently preserved to make recovery possible. Before general availability, we intend to implement and disclose a fixed maximum retention period for those retained failures.
10. Security
We apply technical and organizational measures appropriate to the data, including: encryption in transit (TLS); encryption at rest for sensitive credentials (connector OAuth tokens and any user-provided keys are stored encrypted); least-privilege access; and tenant isolation (each user's data is scoped to their account). No system is perfectly secure; we cannot guarantee absolute security.
11. Your rights (Habeas Data / data subject rights)
Under Colombian Ley 1581 de 2012 (and comparable rights for users elsewhere), you may: access, know, update and rectify your data; request deletion or that we refrain from using it; withdraw the consent you gave; obtain proof of the consent granted; and be informed of how your data is used.
How to exercise them:
- Channel: send your request to hello@stackstitch.dev identifying yourself; we may ask for reasonable proof of identity (and, for organization data, of authorization).
- Information requests (consultas): answered within 10 business days, extendable by up to 5 more with notice.
- Complaints (reclamos) — to correct, update, delete, or for alleged breaches: handled within 15 business days, extendable by up to 8 more with notice.
- Proof of authorization: we keep evidence of the consent/authorization you provided, as the law requires.
- Escalation: if unresolved, you may file a claim with the Superintendencia de Industria y Comercio (SIC), Colombia's data-protection authority, after exhausting the process with us.
You can also disconnect sources, disable product analytics, and delete data through available product controls.
12. Children
The Service is not directed to anyone under 18, and we do not knowingly collect their data.
13. Changes to this policy
We may update this policy; material changes will be notified via the app or stackstitch.dev, and the "Last updated" date will change.
14. Contact
Questions or requests: hello@stackstitch.dev · StackStitch (founding team, Colombia; legal entity pending).